3424
J . Ora. Chem., Pol. 86,No. 10, 1970
DEYRUP AND MOYER
Aziridinemethanols JAMES A. DEYRUPAND CALVINL. MOYER' Department of Chemistry, University of Florida, Gainesville, Florida 39601 Received July 11, 1969 Procedures for the synthesis of substituted aziridinemethanols from aziridine esters have been developed. The stereochemistry of hydride reductions leading to a-phenylaziridinemethanols is highly dependent on the solvent and hydride reagent. Stereochemistry of these hydride reductions is assigned by independent chemical means and discussed in terms of models for asymmetric induction. A number of transformations of aeiridine alcohols in which the aeiridine ring remains intact are described. The mass spectra of aziridinemethanols are also discussed.
Our interest in the interaction of the aziridine ring with incipient positive charge2necessitated the preparation of a series of a-substituted asiridinemethanols and their derivatives. Although there are a number of
SCHEME I
To" N
I
R 2a, R CHB b, R E t-Bu 90% C, R CHZ-Ph
a
'"z% I
t-Bu
H ";",
3
0
R +CH3
known routes to substituted a~iridinernethanols,~ our desire to study a homologous set of compounds and the availability of aziridine esters led t o -developing the high yield sequence of reactions shown in Scheme I. General structures were assigned to these compounds on the basis of their nmr spectra (Table I) as well as infrared, mass spectral (see below), and elemental analyses. TABLE I
N
I
R LiOH
1
I
NMRSPECTRAL PROPERTIES OR t-BUTYLAZIRIDINE DERIVATIVES"
OH
t-Bu
I
4
N
I
t-Bu
0
N
I
LiAIH,
t8u Compd
HI
Hi
Ha
6
7
6% 94% A
B
t-Bu
2.12 0.97 1.67 1.85 0.97 1.33-1.50 1.80 3.43 6 1.83 4.25 0.84 1.37-1.57 7 1.34 1.55 1.97 4.60 0.85 3 2.59 0.77 1.60-1.75 1.03 4 1.82 2.04 2.93 Values are expressed in parts per million (6) downfield from TMS. R = t-Bu.
lbb 2b
(1
It will be noted from Scheme I that a change in the hydride donor and solvent resulted in an inversion of stereoselectivity . Although this result allows investigation of the stereochemical consequences of aziridine interactions with a carbinyl center, an unambiguous stereochemical assignment must precede such study. (1) Support of this research by National Science Foundation Grants GP-5531 and GP-8044 is gratefully aoknowledged. (2) J. A. Deyrup and C. L. Moyer, Tetrahedron Lett., 6179 (1968). (3) (4 R. V. Capeller, R. Griot, M . Haring, and T. Wagner-Jauregg, H e h . Chim. Acta, 40, 1652 (1957); (b) N. H . Cromwell, J. Amer. Chem. Soc., 66, 258 (1947); (c) N. H. Cromwell, J. H . Anglin, Jr., F. W. Olsen, and N. G. Barker, ibid., 1 8 , 2803 (1951); (d) D. K. Wall, J. L. Imbach, A. E. Pohland, R. C. Badger, and N. H . Cromwell, J . Heterocycl. Chem., 6 , 77, 1868.
N _.
NaBH,-MeOH
I
____c
t-Bu
70% 30%
62%
9
Vnrious models have been invoked to explain and predict the stereochemistry of addition to carbonyl groups adjacent to asymmetric carbon atomsa4 I n the absence of interaction between the carbonyl and the asymmetric center, attack apparently occurs on the most stable conformation and from the least hindered side (open chain model). Additions for which this model is valid seldom display high stereoselectivity. Presence of a heteroatom at the asymmetric carbon offers potential coordination of the heteroatom and the carbonyl group with a metal species (cyclic model). Additions to which this model is applicable usually are highly stereoselective. A third model in which the conformational preference is dominated by repulsion between the carbonyl oxygen and heteroatom has been proposed (and c r i t i c i ~ e d ~ ~ ) . (4) (a) D. 5. Cram and D. R. Wilaon, J. Amsr. Chem. Soc., 85, 1245 (1963); (b) G. J. Karabatsos, ibid., 89, 1367 (1967).
J. Org. Chem., Vol. 56, No. 10, 1870
AZIRIDINEMETHANOLS The stereochemistry of reduction of several aziridinyl ketones (8) has recently been studied by Cromwell.ad
8
I n contrast to our results, both LiAlH4 and NaBH4 yielded the same isomer as the major product. The stereochemit.try of this major product was assigned the erythro configuration on the basis of the open chain model. The high stereoselectivity observed in both studies with LiA1H4 is, however, suggestive of the cyclic model which also predicts preferential formation of the erythro isomer. Reductions with NaBH4 were of diminished stereoselectivity in the case of 8 and of inverted stereoselectivity in the case of 5. The different selectivity exhibited by NaBH4 toward 8 and 5 clearly indicates the delicate balance between those factors which govern orientation of addition. Attempts to apply current theories concerning asymmetric induction and reagent size failed to explain convincingly the difference between the two ketones. This failure casts a certain amount of doubt on the validity of stereochemical assignments based solely on these models. Fortunately, it was possible to obtain independent chemical evidence for the stereochemistry of 6 and 7. Reaction of‘ 6 with SOC1, gave oxathiazolidine 9a.5 Similarly, 7 yielded 10. The stereochemistry of these
3425
Chemistry of the Aziridinemethanols. -Our study of the properties of the aziridine ring as a neighboring group necessitated the conversion of these alcohols into esters which would serve as suitable leaving groups., The conventional procedure for tosylate preparations from alcohols using tosyl chloride and pyridine was unsuccessful. This failure presumably arises from the greater base strength of the aziridine (relative to pyridine) and results in open chain products via chloride ion attack on the protonated aziridine ring. Use of triethylamine or sodium hydride as proton scavengers overcame this problem. I n this manner, tosylates of 2a, 2b, and 2c as well as the nosylate of 2b could be prepared in good yield. The sulfonate esters with R = t-Butyl were relatively stable and could be obtained in analytical purity. The other sulfonate esters (R = CHI and CH2Ph) were stable in solution. These solutions could be analyzed spectrally and used for further reactions. Removal of solvent, however, resulted in exothermic polymerization. It appears probable that the t-butyl group inhibits intermolecular alkylation. Attempts to prepare esters of 3, 6 , and 7 have been unsuccessful. I n all cases, conditions required for formation of these derivatives proved too drastic or unselective. The sulfonate esters described above underwent facile displacement in poor ionizing solvents.2 In each case, the displacement product, 11, was obtained in
11
t-Bu
6
(ery t k Po)
9a
J = 2.6-3.6 HZ 7
t-Bu
(threo)
10
compounds can readily be assigned from the magnitude of the coupling constants in these relatively rigid heterocycles. Since oxathiazolidine formation does not affect the configuration a t either asymmetric carbon, the stereochemistry can be assigned as shown. The erythro isomer (6) is, therefore, probably formed by attack from the least hindered side of the carbonyl group as shown in the following structure.
i-Bu (5) J. A. Deyrup, C. L. Moyer, a n d P. S. Dreifus, J. Org. Chem., 36, 3428 (1970).
good yield and uncontaminated by ring-opened or ringexpanded by-products. Although only halide and alcoxide nucleophiles were investigated, it seems certain that such nucleophilic displacements offer quite general procedures for connecting the aziridine methyl system to various groups. Mass Spectra of Aziridinemethano1s.-The closely related series of compounds available from the synthetic work described in this paper prompted us to examine their mass spectral behavior. I n addition to providing precedent for future aziridine mass spectral structural assignments, the mass spectra of these compounds offered the potentially interesting features of interplay between the nitrogen atom and the exocyclic heteroatom. The lower ionization potential of nitrogen relative to oxygen generally results in the predominance of molecular ions formed by the removal of an electron from nitrogen when both nitrogen and oxygen are present.6 I n the case of aziridines, the high s character of the unshared electron pair results in a higher ionization potential (relative to larger nitrogen heterocycle^)^ and thus might diminish the effectiveness of nitrogen in directing fragmentation. I n spite of this fact, it seems quite clear that the important fragmentation pathways can best be rationalized in terms of initial ionization at the nitrogen atom. (6) K. Biemann, “Mess Spectrometry,” McGraw-Hill, New York, N. Y., 1962, pp 87-90; H.Budzikiewicz, C. Djerassi, and D. H. WilIiams, “Mass Spectrometry of Organic Compounds,” Holden-Day, 1967, p p 9-26. (7) H. C. Brown and M. Gerstein, J. Amer. Chem. Soc., 72, 2926 (1950)
J . Org. Chem., Vol. 36,No. 10, 1970
3426
DEYRUP AND MOYER
Although space does not allow depiction of all the mass spectra and discussion of their detailed interpretation, certain common fragmentation patterns and trends dominate the spectra.* These patterns are summarized in Scheme 11. It is reasonable t o expect SCHEME I1
2;
T-7 +N -C-
+N
I
$:
13
I
11,
+
FR) X
yJ I
H
N
15
f-c" I
12
\ /
+N
9
